- Pieces of data management plans should not be copied from others. Making a data management plan is like following a checklist: you need to think about different aspects of your own project, and become aware of potential data management issues before they become data management problems. Copying an ideal backup policy from a sister project does not help this cause.
- Data management plans should support production of F, A, I, R data (Findable, Accessible, Interoperable, Reusable). This includes activities during the project as well as archiving after the project.
- Data management when done well (which includes the "Active" in "Active Data Management Plans") helps scientists perform their work better. To convince scientists, it is important that the different sections in a data management plan template/questionnare motivate what they do to improve science and reduce risks.
- To tell scientists that they must do data management to prove that they are trustworhy is counter-productive. The main objective is not to satisfy the data management police, but for better science.
- DMP tools should explicitly support the updating of the plans during the project.
- It is good if the execution of a DMP is done by a WP leader that is made responsible for a good outcome, and monitored from a distance by the funding agencies.
- There are so many aspects to data management, between the hard-core ICT and the subject matter on hand and every combination level in between, that it is impossible that a single person would be able to be an expert on everything. I am working on a (woofully incomplete) mind map of DMP issues for life science projects, to print out as A0 you need 8 point print. However, the Dunning-Kruger effect makes that many people consider themselves data-experts and do not call for help.
- In my experience judging DMP's and advising scientists about how to execute them well, I have seen that a single question that is not applicable can make people tune out and not take the rest of the questionnare seriously.
- There are only very few questions/answers in a DMP that actually can be judged. For most questions (especially those that are present to increase awareness of potential issues) any thoughtful answer is good.
There doesn't seem to be an obvious way to comment on a wiki page, so I'm just going to add some thoughts below yours.
This is a great set of points to add to discussion. I particularly echo 1, 8 & 9.
At the moment DMPs are too driven by compliance. We need to focus more on the benefits to researchers of taking time to reflect on potential problems before they arise. As you say, an answer that's considered and shows reflection is what we should be looking for.
We have also encountered examples where researchers completely disengage if a single question is irrelevant to their context, so I think we need to try to discourage people from asking lots of questions that aren't well tailored. There's a tendency for organisations to make up new templates from scratch, but we should be learning more from organisations that have been doing this for a while and have taken time to refine their requirements and approach. More thought should also be given to how each question directly benefits the researcher, for example by helping them to establish procedures or by generating service requests etc.
These are great points to explore further - hope it's a useful discussion
Comment added by Sarah Jones, Digital Curation Centre, on 24th September 2015, 08:38